Reconfigurable cognitive networks

Switches are the fundamental elements of intelligence that enable logic creation and reasoning. In the past, the idea of using weights to control the switches has been used for emulating cognitive functions. In this research, we present a resistive memory based switch matrix architecture that can be used for implementing different cognitive functions. A forward path algorithm is proposed to implement circuit architectures in the proposed re-configurable switch matrix. Re-configurable capability of the switch matrix provides reusability and standardization. In comparison with conventional logic circuits, the proposed circuit indicate lower power, lower time delay and use fewer number of components.